# Genome wide association study of response to interval and continuous exercise training: the Predict-HIIT study

PMID: 33985508
Journal: Journal of Biomedical Science
Published: 2021-05-13
Authors: Williams CJ, Li Z, Harvey N, Lea RA, Gurd BJ, Bonafiglia JT, Papadimitriou I, Jacques M, Croci I, Stensvold D, Wisloff U, Taylor JL, Gajanand T, Cox ER, Ramos JS, Fassett RG, Little JP, Francois ME, Hearon CM Jr, Sarma S, Janssen SLJE, Van Craenenbroeck EM, Beckers P, Cornelissen VA, Howden EJ, Keating SE, Yan X, Bishop DJ, Bye A, Haupt LM, Griffiths LR, Ashton KJ, Brown MA, Torquati L, Eynon N, Coombes JS

## Question

Can genome-wide genetic variation predict individual VO2peak response to HIIT, sprint interval training, or moderate continuous training?

## Summary

This genome-wide association study tested whether common genetic variants could predict VO2peak response to interval or continuous exercise training. In 507 participants from 18 exercise interventions, no genome-wide significant predictors were found, and a 12-SNP predictor did not successfully predict response in cross-validation or an independent HIIT validation cohort.

## Population

- Predict-HIIT exercise intervention participants with GWAS data; validation in sedentary apparently healthy adults.
- Sample size: 507
- Age: Predict-HIIT 55.9 +/- 16.9 years; Improve-HIIT 31.6 +/- 9.8 years
- Sex: Predict-HIIT 385 male/122 female; Improve-HIIT 9 male/30 female
- Fitness level: Mixed; validation cohort sedentary
- Health status: Predict-HIIT mostly elderly or clinical/pathology samples; validation apparently healthy

## Methodology

- Multi-centre genome-wide association and validation study using exercise intervention datasets
- Varied across 18 included interventions; validation cohort trained for 6 weeks
- Multi-centre exercise intervention data analysis
- Controlled study design.

## Protocol

- High-volume HIIT category.
- Modality: Varied exercise interventions.
- Work intervals: Varied; validation used 4-minute intervals.
- Recovery: Varied; validation used 3-minute recoveries.
- Sets or repetitions: Varied; validation used 4 intervals.
- Intensity: High-volume HIIT category had >=15 min high-intensity work; validation targeted 90-95% HRmax.
- Frequency: Varied; validation 3 sessions/week.
- Program length: Varied; validation 6 weeks.
- MICT and low-volume HIIT/SIT categories.
- Modality: Varied exercise interventions.
- Work intervals: Low-volume HIIT/SIT <15 min high-intensity work; MICT continuous exercise.
- Intensity: MICT 64-76% HRmax; low-volume HIIT/SIT varied.
- Session duration: MICT at least 30 min continuous exercise.

## Outcomes

### GWAS loci
Status: no clear change
No covariate-adjusted variants reached genome-wide significance.

12 suggestive loci p<1e-5; strongest near MAGI2 rs6959961 p=2.61e-7.

### Predictor score
Status: no clear change
The 12-SNP predictor did not predict VO2peak response in cross-validation or independent validation.

Validation analyses p>0.1.

### VO2peak response
Status: mixed
VO2peak response varied substantially, with many participants classified as likely nonresponders or uncertain responders.

Predict-HIIT overall response 3.0 +/- 3.8; likely nonresponders 41.4%, likely responders 24.1%, uncertain 34.5%.

## Practical Insights

- Do not promise genotype-based personalization of HIIT response from current evidence.
- Individual VO2peak response varies, so programs should measure user progress rather than assume a uniform response.
- The validation 4 x 4-minute protocol is reproducible, but the paper's main contribution is genetics rather than programming.

## Limitations

- Underpowered for genome-wide discovery of small genetic effects.
- Heterogeneous source interventions and participant populations.
- Mostly European-descent data.
- Adherence and safety details were not the focus of the paper.

## Safety And Adherence

- No adverse events were reported in this secondary GWAS article.
- Validation HIIT was supervised, but detailed safety and adherence reporting were not central to this paper.
- Ethics approvals were reported for both Predict-HIIT and Improve-HIIT.

## Original Sources

- [PubMed](https://pubmed.ncbi.nlm.nih.gov/33985508/) (pubmed)
- [DOI](https://doi.org/10.1186/s12929-021-00733-7) (doi)
- [PMC full text](https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8117553/) (full text)

## Agent Guidance

Preserve the paper-level scope of this note. Do not generalize beyond the population, protocol, measured outcomes, and limitations above.